Title | ||
---|---|---|
Discriminative and semantic feature selection for place recognition towards dynamic environments |
Abstract | ||
---|---|---|
•A FCN model is proposed to estimate an activation map which can indicate the stability of regions for features detection.•A hybrid supervision for training is introduced that combine strong supervised regression and self-supervised optimization.•Selected features are experimentally concluded to be more robust in BoW-based place recognition and overall SLAM system. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1016/j.patrec.2021.11.014 | Pattern Recognition Letters |
Keywords | DocType | Volume |
Place recognition,Feature points,SLAM,Deep learning | Journal | 153 |
ISSN | Citations | PageRank |
0167-8655 | 0 | 0.34 |
References | Authors | |
0 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yuxin Tian | 1 | 2 | 4.77 |
Jinyu Miao | 2 | 0 | 2.03 |
Xingming Wu | 3 | 43 | 13.16 |
Haosong Yue | 4 | 2 | 3.10 |
Zhong Liu | 5 | 0 | 0.34 |
Weihai Chen | 6 | 1 | 1.74 |